1. A Fast Method for Computing High-Significance Disease Association in Large Population-Based Studies 
Gad Kimmel, Ron Shamir 
The American Journal of Human Genetics 
Volume 79, Issue 3, Pages (September 2006)
DOI: /507317
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

2. Figure 1
Comparison of running times of the two algorithms that test the disease association of individual SNPs. We present run times of RAT (×) and SPT (circles) on simulated data under the coalescent model with recombination. The target P value was 10-6 in all cases. Running times reflect savings due to importance sampling only, without the additional possible savings due to LD decay. The Y-axis gives the logarithm (base 10) of the running time in seconds.
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

3. Figure 2
Convergence of RAT to the “true”P value. Each of the five figures represents a different experiment with 100 controls and 100 cases of simulated SNPs in a 1-Mb region (∼3,000 SNPs), under the coalescent model. SPT P value was evaluated by applying 10,000 (A, D, and E) or 100,000 (B and C) permutations. The horizontal dashed lines correspond to the 95% CI of SPT P value. Each graph corresponds to the RAT P value.
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

4. Figure 3
Dependence of accuracy on the P value. Data sets were simulated SNPs under the coalescent model with recombination of a 1-Mb region. To obtain different P values, we performed the simulations with different numbers of cases and controls ranging from 50 to 500.
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

5. Figure 4
Running times of RAT and SPT at different P values. The data sets are simulated data under the coalescent model with recombination of a 1-Mb region (∼3,300 SNPs) of 5,000 cases and 5,000 controls. To obtain different P values, the simulations were performed with different phenocopy rates (λ parameter) of the multiplicative disease model. × = RAT; circles = SPT. The Y-axis shows the logarithm (base 10) of the running time in seconds, and the X-axis shows the logarithm (base 10) of the P value.
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

6. Figure 5
Effect of the linkage upper bound used on the P value calculated by RAT. Data sets A–D are the first 10,000 SNPs in chromosomes 1–4, respectively, of 200 cases and 200 controls, which were amplified from 60 unrelated individuals (the CEPH population from the HapMap project). The dashed lines correspond to the 95% CI of the calculated P value. The wide range of P values obtained is probably due to the random choice of the disease SNP, the stochastic model of the disease, and chromosomal characteristics.
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions

7. Figure 6
Effect of the LD decay on the speed of RAT. The Y-axis shows the time required by the permutation phase of the RAT algorithm. The X-axis shows the assumed linkage bound. The data are the first 10,000 SNPs in chromosome 1 of 1,000 cases and 1,000 controls, which were amplified from 60 unrelated individuals (the CEPH population from the HapMap project).
The American Journal of Human Genetics  , DOI: ( /507317)
Copyright © 2006 The American Society of Human Genetics Terms and Conditions